Unmasking the Annexin I Interaction from the Structure of Apo-S100A11

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Anne C. Dempsey,1 Michael P. Walsh,2 some members of this family may act as calcium signal- and Gary S. Shaw1,* ing or sensor proteins (Drohat et al., 1998; Maler et al., 1Department of Biochemistry 2002; Otterbein et al., 2002; Sastry et al., 1998; Smith The University of Western Ontario and Shaw, 1998b). The S100 proteins are small acidic London, Ontario N6A 5C1 proteins of approximately 10–12 kDa in molecular weight Canada and found only in vertebrates. To date there have been 2 Department of Biochemistry and Molecular nineteen S100 proteins identified, most of which appear Biology to be cell specific (Donato, 2001; Heizmann, 1999; Heiz- University of Calgary mann et al., 2002). For example, S100B is concentrated Calgary, Alberta T2N 4N1 in glial cells (Petrova et al., 2000), while S100P is located Canada mainly in placental tissue (Becker et al., 1992). Several of the S100 proteins have also been linked with disease states such as Alzheimer’s and rheumatoid arthritis, Summary usually via the overexpression of the protein (Odink et al., 1987; Van Eldik and Griffin, 1994). However, the cause S100A11 is a homodimeric EF-hand calcium binding of the overexpression and details of the roles of these protein that undergoes a calcium-induced conforma- proteins in disease is still uncertain. Other S100 proteins tional change and interacts with the phospholipid have been shown to interact with a variety of proteins binding protein annexin I to coordinate membrane as- such as tubulin (Garbuglia et al., 1999), annexins I, II, sociation. In this work, the solution structure of apo- and VI (Bianchi et al., 1992; Garbuglia et al., 1998; Gerke, S100A11 has been determined by NMR spectroscopy 1990; Seemann et al., 1996), p53 (Baudier et al., 1992; to uncover the details of its calcium-induced structural Delphin et al., 1999), and actin (Fujii et al., 1990), and change. Apo-S100A11 forms a tight globular structure are proposed to regulate processes such as assembly, having a near antiparallel orientation of helices III and phosphorylation, and membrane interaction. IV in calcium binding site II. Further, helices I and IV, S100A11 (S100C, calgizzarin) is a member of the S100 and I and I؅, form a more closed arrangement than family, which is often expressed in smooth muscle (Allen observed in other apo-S100 proteins. This helix arrange- et al., 1996; Scho¨ nekess and Walsh, 1997; Seemann et ment in apo-S100A11 partially buries residues in heli- al., 1996). S100A11 has been shown to interact with the ces I (P3, E11, A15), III (V55, R58, M59), and IV (A86, phospholipid binding protein annexin I in a calcium- C87, S90) and the linker (A45, F46), which are required sensitive manner (Seemann et al., 1996). In this mecha- for interaction with annexin I in the calcium-bound nism, the N-terminal helix of annexin I is released from state. In apo-S100A11, this results in a “masked” bind- interaction with the third repeat in the annexin molecule ing surface that prevents annexin I binding but is un- (Gerke and Moss, 2002). This helix then associates with covered upon calcium binding. calcium-bound S100A11 where it could act as a bridge to link pairs of phospholipid-containing membranes. In the absence of calcium, annexin I has negligible affinity Introduction for S100A11 or the phospholipid matrix. Consistent with this role and an intracellular location near the interior of The calcium ion acts as an important second messenger membranes, the calcium affinity for S100A11 is lower controlling integral cellular events such as neurotrans- Ca Ϫ4 M) than that typically found for other S100 10ف Kd) mitter release, vision, and muscle contraction. In many Ϫ Ϫ M), because the local calcium 6 10–5 10ف proteins (K Ca cases, a member of the EF-hand family of calcium bind- d concentration appears higher near the membrane. It is ing proteins plays an intermediary role in these pro- likely the affinity of S100A11 for calcium is increased in cesses through either calcium-dependent interactions the presence of a target protein, as has been noted for with selected target proteins or maintenance of intra- calmodulin (Martin et al., 1996; Olwin and Storm, 1985; cellular calcium levels. In the first mechanism, calcium Yazawa et al., 1987). In support of this hypothesis, the fold in-10ف binding to an EF-hand protein results in a significant affinity of S100A11 for calcium is enhanced conformational change in the protein and a change in the presence of the hydrophobic probe 2-p-toluidinyl- its surface properties, facilitating the interaction with naphthalene-6-sulfonate (Allen et al., 1996). Other stud- other proteins. The EF-hand proteins troponin-C and ies have shown that S100A11 associates with F-actin, calmodulin are members of this category of “sensor” where it can inhibit its myosin Mg2ϩ-ATPase activity proteins (Ikura, 1996), having their respective interac- (Zhao et al., 2000). tions with troponin-I (McKay et al., 1997) and myosin S100A11, like most S100 proteins, is a dimeric protein, light chain kinase (among others; Ikura et al., 1992) con- with each monomer comprising two helix-loop-helix mo- trolled by calcium binding. tifs that coordinate the calcium ions. Calcium binding Recently, three-dimensional structures of proteins in site I is comprised of helices I and II and a 14 residue the S100 family of EF-hand proteins have indicated “pseudo” calcium binding loop, while site II contains helices III and IV with a 12 residue canonical calcium *Correspondence: [email protected] binding loop. Calcium binding to most S100 proteins Structure 888

Figure 1. Selected Regions of 600 MHz NMR Spectra Used for the Backbone Assignment of Apo-S100A11 The spectra show 15N planes for the residues N40–A44 at the C terminus of helix II and start of the linker. For each pair of planes, the CBCA(CO)NH is shown on the left and the HNCACB is shown on the right. The C␣ and C␤ for the indicated residue are shown on the HNCACB spectra and the corresponding C␣ and C␤ for the previous (i Ϫ 1) residue are shown in the CBCA(CO)NH. The spectrum was collected on a

1.0 mM apo-S100A11 sample in 90% H2O/10% D2O containing 50 mM KCl, 10 mM dithiothreitol (pH 7.25) at 35ЊC. causes a significant reorientation of one or more of the are facilitated by an open conformation of helices III helices, resulting in an exposure of hydrophobic resi- and IV, which are near perpendicular to one another. dues previously buried in the apo-state (Maler et al., Comparison to other apo-S100 protein structures sug- 2002; Smith and Shaw, 1998a). For example, calcium gests that these helices have reoriented upon calcium binding to human S100B triggers a reorientation of helix binding in order to create a hydrophobic cleft for target III with respect to helix IV and extends the ␣ helix in binding. However, the absence of an S100A11 structure helix IV by one turn (Drohat et al., 1998; Matsumura et in the calcium-free state and the differences from the al., 1998; Smith and Shaw, 1998b). This results in the calcium-bound S100B complexes makes the details of exposure of several hydrophobic residues in the C termi- the conformational change and protein-protein interac- nus and linker region of the protein, providing a surface tions in S100A11 difficult to assess. In this work, we for target interaction. In S100A6, calcium binding in- have determined the three-dimensional structure of rab- duces a less dramatic conformational change, although bit apo-S100A11 by NMR spectroscopy in order to iden- a hydrophobic surface is still exposed. In general, it is tify the conformational change that occurs in this protein thought that differences in the hydrophobic surfaces and identify residues that confer specificity for target between S100 proteins and the magnitude of the confor- protein binding. The structure reveals that helices I and mational change confer target specificity for a variety IЈ at the dimer interface of apo-S100A11 form a more of different proteins. closed arrangement compared to other apo-S100 pro- Recently, the crystal structure of calcium-bound teins, while helices III and IV form a near antiparallel S100A11 in complex with an N-terminal peptide of annexin arrangement. This conformation of apo-S100A11 buries I has been determined (Rety et al., 2000). The structure residues required for annexin I interaction found in heli- shows that the annexin interface occurs through interac- ces I, III, and IV and the linker region. Upon calcium tions with the linker and helices III and IV of one monomer, binding to S100A11, these residues are unmasked to and helix I of the other monomer in S100A11. Thus, the reveal a stunning contiguous binding surface. annexin molecule “bridges” the two S100 monomers. This interaction is very different from the interactions of Results and Discussion another S100 protein, S100B, with two of its targets, p53 (Rustandi et al., 2000) and a 12 residue synthetic The backbone and side chain resonance assignments peptide, TRTK-12 (McClintock and Shaw, 2003). In both for apo-S100A11 (Rintala et al., 2002) were determined of these cases, there are no interactions with helix I in by the collection and analysis of a set of two- and three- S100B, and each target peptide interacts with only one dimensional NMR experiments. The NMR spectra of monomer. In the calcium-bound S100A11/annexin I apo-S100A11 were of high quality and well resolved. complex, interactions between S100A11 and annexin I Figure 1 shows a strip plot of the CBCA(CO)NH and Structure of Apo-S100A11 889

Figure 2. Structures of Apo-S100A11 Stereo view of the backbone superposition for the family of 19 low-energy structures of apo-S100A11 based on nOe, dihedral, and chemical shift NMR restraints. The two S100A11 monomers in the protein are colored magenta and blue to distinguish them. Helices are labeled I–IV in the first monomer and IЈ–IVЈ in the second. The superposition utilized the residues found in the regular secondary structure of the protein (eight helices; 5–20, 32–38, 56–60, and 74–85 and 5Ј–20Ј,32Ј–38Ј,56Ј–60Ј, and 74Ј–85Ј) and had an rmsd to the mean structure of 0.45 Ϯ 0.08 A˚ for backbone residues and 1.15 Ϯ 0.21 A˚ for all heavy atoms, respectively.

HNCACB spectra for several sequential residues (N40– Description of the Structure of Apo-S100A11 A44) in the linker region of the protein. There were some A total of 3028 distance restraints and 192 dihedral re- difficulties assigning the backbone of the serine residues straints was used for the calculation of the three-dimen- of the protein. To make their assignment possible, specifi- sional structure of the apo-S100A11 dimer. Figure 2 cally 15N-glycine/serine-labeled S100A11 was prepared shows a superposition of an ensemble of 19 structures and a 1H-15N HSQC of those residues was collected of apo-S100A11. These structures were chosen based and analyzed. This also facilitated the assignment of on their low energy and contained no distance violations cysteine residues C9 and C87, because a small amount greater than 0.5 A˚ and no angle violations greater than of transamination occurred during the labeling step. The 5Њ. The family of structures shows the presence of four chemical shifts for C␤ of these cysteine residues indi- well-defined helices (I–IV) within each monomer. The cated the protein was in a fully reduced state. This ob- rmsd of these regions in the dimer relative to the mean servation is in contrast to that observed in the structure structure is 0.45 Ϯ 0.08 A˚ for the backbone and 1.15 Ϯ of calcium-bound S100A11 in complex with the annexin I 0.21 A˚ for all heavy atoms. Each of the four helices is peptide, where a C9-C9Ј intermolecular disulfide is pres- well defined (average 0.20 Ϯ 0.08 A˚ ), with helix IV having ent (Rety et al., 2000). Although apo-S100A11 also had the lowest rmsd (0.14 Ϯ 0.06 A˚ ), and helix II being the a strong tendency to oxidize, this was suppressed using least well defined (0.23 Ϯ 0.10 A˚ ). A short antiparallel high quantities of dithiothreitol in all NMR samples. ␤ sheet was identified within each monomer between Stereo-specific assignment of the methyl groups of va- residues T29 and S31 in calcium binding loop I, and line and leucine residues was achieved by the analysis residues Q70 and D72 in loop II. For this region, nOes of a 13C-coupled 1H-13C HSQC spectrum of fractionally were found across the ␤ sheet between protons of T29 13C-labeled S100A11 (Neri et al., 1989). This method dif- with L71 and D72, L30 with L71, and S31 with Q70. The ferentiates the pro-R methyl groups (H␥1 or H␦1 of valine extreme N terminus and the last 15 residues of the C and leucine, respectively) from the pro-S methyl groups terminus of apo-S100A11 are less well defined in the based on 13C-13C coupling. In apo-S100A11, all 5 valine structures. This is likely a result of the lack of long-range and 11 leucine residues had stereo-specific assign- nOes between S1–P3 and A86–F101, indicating these ments completed. regions are flexible in solution. Further, chemical shift Structure 890

Table 1. Structure Statistics for Apo-S100A11 Structure Statisticsa Final Energies (kcal/mol) Total energy 159.42 Ϯ 59.13 NOE energy 94.95 Ϯ 7.71 Lennard-Jones VDW energy Ϫ191.93 Ϯ 45.57 Distance restraints Ͼ 0.5 A˚ 0 Angle violations Ͼ 5Њ 0 Rmsd from experimental distance restraints (A˚ ) 0.0204 Ϯ 0.0008 Rmsd from experimental angular restraints (A˚ ) 0.5825 Ϯ 0.0596 Restraints for Structural Calculations Total NOE Restraints 2964 Intraresidue 1336 Sequential 658 Short-range 594 Long-range 264 Intermolecular 112 Dihedral bond restraints 192 Hydrogen bonds 32 Ramachandran Plot Residues in favorable regionsb 89.3% Coordinate Precisionc Rmsd to Mean (A˚ ) Backbone atoms 0.45 Ϯ 0.08 Heavy atoms 1.15 Ϯ 0.21 a All statistics were calculated for a family of 19 structures. b This reflects residues in both most favored and additionally favored regions. c Precision is calculated for residues found in helices I (5–20), II (32–38), III (56–60), and IV (74–85) in apo-S100A11.

data were unable to find a consistent solution for φ,␺ to that observed for apo-S100A1 (⍀ϭ120Њ) and apo- angles in these regions, indicating a range of angles is S100A4 (⍀ϭ119Њ) (Rustandi et al., 2002; Vallely et al., likely sampled for these residues. Table 1 shows the 2002). Most interestingly, the interhelical angle between structural statistics for this family of structures. Analysis helices III and IV was found to be positive (⍀ϭ154Њ) of this family of structures indicated that 89.3% of all as in bovine apo-S100B (⍀ϭ164Њ) (Kilby et al., 1996), of the residues are in the most favorable and additionally apo-S100A6 (⍀ϭ150Њ) (Maler et al., 1999; Otterbein et favorable regions of the Ramachandran plot. al., 2002; Potts et al., 1995), and apo-S100A4 (⍀ϭ162Њ) The four helices in each S100A11 monomer were se- (Vallely et al., 2002), while it has been found to be a lected based on the superposition of structures and φ,␺ negative orientation in rat apo-S100B (Drohat et al., angles from the Ramachandran plot and comprise helix I 1999), apo-S100A1 (Rustandi et al., 2002), and apo- (E5–Y20), helix II (K32–F38), helix III (L56–M60), and helix S100A3 (Fritz et al., 2002). This interaction in apo-S100A11 IV (Q74–V85). The two ␤ strands consist of the residues was supported by nOes from M59 to L79, M59 to L83, and T29–S31 and Q70–D72 arranged in an antiparallel ␤ M60 to L79. Together, these observations for calcium sheet. Table 2 describes the interhelical angles of apo- binding sites I and II indicate the helix interactions in S100A11. The arrangement of helices I and II for calcium apo-S100A11 are most reminiscent of apo-S100A4. binding site I in apo-S100A11 (⍀ϭ116Њ) is most similar Dimerization of apo-S100A11 occurs through interac-

Table 2. Interhelical Anglesa for Apo-S100A11 and Calcium-Bound S100A11/Annexin I Complex Helix Pair Apo-S100A11b S100A11-An Ib,c Differenced Rmsde I-II 116 Ϯ 3 129 13 1.00 I-III Ϫ73 Ϯ 5 Ϫ112 Ϫ39 1.62 I-IV 115 Ϯ 2 127 12 1.18 II-III 159 Ϯ 5 118 Ϫ41 1.38 II-IV Ϫ41 Ϯ 3 Ϫ29 12 1.49 III-IV 154 Ϯ 3 115 Ϫ39 2.65 I-IЈϪ137 Ϯ 3 Ϫ154 Ϫ17 - IV-IVЈ 153 Ϯ 5 153 0 - a Determined using the program iha.1.4 (S. Gagne´ , Universite´ Laval). b Helices were defined as residues 5–20, 32–38, 56–60, and 74–85 in apo-S100A11; 7–22, 34–40, 56–60, and 76–87 in the S100A11/annexin I complex. c Calcium-bound S100A11 in complex with the N-terminal helix from annexin I (Rety et al., 2000); PDB ID code 1QLS. d The difference reflects the change in interhelical angle between apo-S100A11 and the S100A11/annexin I complex. A positive number indicates a clockwise change and a negative number indicates a change in a counter-clockwise direction. e The backbone rmsd determined from a superposition of helices as defined in (b). Structure of Apo-S100A11 891

Figure 3. Identification of Residues at the Dimer Interface of Apo-S100A11 13 15 13 13 1 Regions of the C, N NOESY (A) and CF1-filtered, F3-edited NOESY (B) showing the C plane at 23.95 ppm where L43 ␦2 CH3 is located at a H frequency of 0.56 ppm (shown in [A] by the vertical line). The 13C,15N NOESY (A) shows nOe crosspeaks from all protons whose attached 13Cis 13 found on this plane in apo-S100A11. The CF1-filtered, F3-edited NOESY spectrum (B) shows only those nOes (indicated by dashed lines) across the dimer interface involving L43 ␦2 CH3 from the linker on one monomer of apo-S100A11 and T6, E7, and I10 of helix IЈ of its partner monomer. tions between helices I, IЈ, IV, and IVЈ, which form an protein (McClintock and Shaw, 2003; Rety et al., 1999, X-type bundle. Specifically, the interhelical crossing 2000; Rustandi et al., 2000). Of these, only S100B has angles for helices IV with IVЈ (153 Ϯ 5Њ) was similar to structures available for the apo- (Drohat et al., 1999; that observed in calcium-saturated S100A11 in complex Kilby et al., 1996), calcium- (Drohat et al., 1998; Matsu- with the annexin I peptide (Rety et al., 2000), while helices mura et al., 1998; Smith and Shaw, 1998b), and target- IandIЈ (137 Ϯ 3Њ) were more closed than in the complex bound states (McClintock and Shaw, 2003; Rustandi et (Ϫ154Њ). Residues that define the dimer interface for apo- al., 2000). For S100A11, a three-dimensional structure 13 S100A11 were identified from a CF1-filtered, F3-edited is available for the protein bound to the N-terminal region NOESY spectrum of the protein where one monomer was of the phospholipid binding protein, annexin I (Rety et 13 15 unlabeled and the other was C, N-labeled. Figure 3 al., 2000). Thus, the current structure of apo-S100A11 13 ␦ shows the C plane for L43 2 CH3 (23.95 ppm) of the allows a comparison with the bound structure to deter- 13 resulting CF1-filtered, F3-edited NOESY spectrum and mine regions that undergo a conformational change and 13 13 15 the corresponding C plane of the C, N-NOESY spec- are important for target protein recognition. Figure 4 trum. The spectra show that residue L43 in the linker shows a comparison of the structures of apo-S100A11 makes several contacts across the dimer interface with and the complex of calcium-bound S100A11 with an residues T6, E7, and I10 in helix IЈ. Other nOes from these N-terminal peptide of annexin I. The structures reveal data showed that T6 in helix I had close contacts with that every helix pair in S100A11, except that of helices L13Ј and V16Ј in helix IЈ, while C9 and I10 in helix I were IV and IVЈ, shifts upon calcium binding and interaction close in space to L13Ј. Further, T6 and I10 in helix I had with annexin I (Table 2). One of the most conservative nOes with G80Ј and A84Ј in helix IVЈ. The helix IV-IVЈ interface was defined by nOes of F73 with G81Ј, A84Ј, changes in helix interaction involves helices I and II, and V85Ј as well as nOes from L77 to G80Ј, G81Ј, and which surround calcium binding site I. The interhelical A84Ј. These interactions aligned helices I and IV in oppo- angle for these helices is more open in the annexin Њ site directions from helices IЈ and IVЈ, respectively, in complex by approximately 13 . Superposition of helices the dimer. I and II from apo-S100A11 with those in the annexin complex reveals only a 1.00 A˚ rmsd between the two. Apo-S100A11 Global Changes upon This indicates that the conformation of site I is altered Complexation with Calcium and Annexin I very little by calcium and/or annexin I binding. Similar There have only been four three-dimensional structures observations have been made for S100B (Smith and of an S100 protein in complex with calcium and a target Shaw, 1998b), S100A6 (Maler et al., 2002; Otterbein et Structure 892

Figure 4. Global Conformational Change in S100A11 The ribbon diagrams of apo-S100A11 (A) and calcium-bound S100A11 (B) complexed to the N-terminal annexin I peptide (Rety et al., 2000) are shown in similar orientations. Although the structures are symmetric, only one monomer from each dimer is colored for easier viewing. For calcium-bound S100A11 (B), the annexin I peptide (magenta) interacts with helices III and IV and the linker of one monomer and helix IЈ of the partner monomer. The structures show that the orientations of helices I (purple) and II (yellow) in site I are similar upon calcium binding and complex formation. Helix I has shifted by approximately 17Њ, opening up the helix I-IV interhelical interaction. Helix III is nearly parallel (⍀ϭ154 Ϯ 4Њ) to helix IV in apo-S100A11 and is almost perpendicular (⍀ϭ115Њ) in the calcium-bound structure. Helix IV is notably shorter in apo-S100A11, ending near residue A85, whereas this helix is 10 residues longer in the annexin I complex. The calcium-bound S100A11 structure also contains a short helix in the linker region (gray) found in some structures of apo-S100A11.

al., 2002; Sastry et al., 1998), and calbindin D9k (Akke et One result of the reorientation of helices I and IV in apo- al., 1992; Kordel et al., 1993; Skelton et al., 1995; Wim- S100A11 is a subtle redistribution of side chain exposure berly et al., 1995), which all possess a similar 14 residue of several residues in both helices. Figure 5 shows that pseudo-EF-hand calcium binding site I and is consistent the pattern of accessible surface area for apo-S100A11 with this loop being in a “calcium-ready” state. and S100A11 in complex with calcium and the annexin Whereas the helix I-II interaction in apo-S100A11 ap- peptide are different for several residues in these re- pears largely conserved upon calcium binding, it is clear gions. In apo-S100A11, residues P3, I10, E11, I14, and that the relationship of helix II to helices III and IV is A15 increase their side chain exposure by as much as altered. This is not obvious from the small interhelical 25% upon calcium binding. Similarly in helix IV, residues angle shift between helices II and IV, which only differs N78, G81, G82, L83, V85, A86, and C87 all increase their by 12Њ in the two structures (Table 2) but is manifested side chain exposure to solvent. As will be described, in a larger rmsd between the helix pairs of 1.49 A˚ .A many of these residues are critical for annexin I binding. detailed analysis of S100A6 has revealed that similar The largest structural changes upon calcium binding changes in the helix II position result from the repacking and target protein interaction involve helices III and IV. of several residues in this helix upon calcium binding This results in extreme changes in side chain surface (Maler et al., 2002). Indeed, residues near the C terminus accessibility (Figure 5) in the calcium binding loop resi- of helix II (E34, S37, N40-E42) are more exposed in apo- dues for site II (D64–E75). Similarly large changes are S100A11 than in the calcium form of the protein (Figure noted for site I (A21–E34). The position of helix III 5). This switch in residue packing may be a consequence changes by approximately 40Њ with regard to each of of the increased exposure of residues in the adjacent helices I, II, and IV. In combination with the helix I-IV linker and helix III that are involved in target protein and I-IЈ reorientation, this has the dramatic effect of binding. opening up one face on each side of the apo-S100A11 In apo-S100A11, helices I and IV are closer to a per- molecule (Figure 6) and placing helix III nearly perpen- pendicular orientation (⍀ϭ115Њ) than in the presence dicular to helix IV (⍀ϭ115Њ) in the calcium-bound struc- of calcium and the annexin peptide (⍀ϭ127Њ). The ture (Figure 4). magnitude of this change is nearly twice that observed The second most notable feature in apo-S100A11 is upon calcium binding to S100B (Drohat et al., 1998; the absence of helix between A86 and F101 (Figure Matsumura et al., 1998; Smith and Shaw, 1998b). Fur- 4). Upon calcium (and annexin I) binding, this region ther, a comparison of the changes in the intramolecular becomes helical extending to T95, the third to last resi- I-IV angle with those of helices I-IЈ (17Њ) and IV-IVЈ (0Њ) due fitted in the X-ray structure of the complex (Rety reveals it is the movement of helix I that is most responsi- et al., 2000). In S100A11, it is not known whether this ble. It is possible that a disulfide link between C9 and induction of helix is a result of calcium or peptide bind- C9Ј in the S100A11-annexin I structure (Rety et al., 2000) ing. However, a similar observation has been made for may contribute to the interactions of helices I and IV. human and bovine S100B upon calcium binding alone However, C9 is buried in both structures (although not and results in the extension of helix IV by approximately in an oxidized state in apo-S100A11), suggesting a re- one turn (Matsumura et al., 1998; Smith and Shaw, arrangement of this residue has not occurred (Figure 5). 1998b). For S100A11, mutational studies have shown Structure of Apo-S100A11 893

Figure 5. Unmasking of Buried Residues in Apo-S100A11 Important for Annexin I Interactions in the Calcium-Bound S100A11/Annexin I Complex The fractional accessible surface areas for the side chains of residues in apo-S100A11 (A) and calcium-bound S100A11 (B) (Rety et al., 2000) were calculated from their three-dimensional structures using the program VADAR (University of Alberta). The accessible surface area for the calcium-bound S100A11 complex was determined after removal of the peptide coordinates from the structure. The bottom panel (C) shows the difference in fractional accessible surface area between the two structures and corresponds to the surface area for side chains in apo- S100A11 that become exposed in the calcium-bound annexin I complex. Bars colored dark blue represent residues in S100A11 that have contact (Ͻ5A˚ ) with the annexin I peptide. Many of these increase in side chain exposure by greater than 20% upon calcium binding. Cyan- colored bars represent residues in apo-S100A11 that increase their side chain exposure by Ͼ20% but do not have direct contact with annexin I in the calcium-bound complex. To facilitate comparison, all sequence numbers correspond to those of rabbit S100A11. that deletion of residues D91–I94 (E89–V92 in rabbit) less available for interaction with the N-terminal helix in dramatically alters the ability of the protein to interact annexin I. Further, several of these residues, important with the N-terminal annexin I peptide upon calcium bind- for interaction with annexin I, do not form a contiguous ing (Seemann et al., 1996). Because these residues are surface required for peptide binding (Figure 6). Specifi- not in a helical structure in apo-S100A11 this would cally, residues P3, I10, E11, S12, I14, and A15 in helix I suggest that the conformational change in this region are more buried in apo-S100A11 than in the S100A11/ upon calcium binding is important for interaction with annexin I complex (Figures 5 and 6). Similar observations annexin I. are made for helices III and IV where residues V55, R58, and V59, and L83, A86, C87, and S90, respectively, are Specificity for Annexin I Target Interaction more buried in apo-S100A11 but have protein-peptide The structure of apo-S100A11 reveals that many resi- interactions in the complex. These observations clearly dues in helices I, III, and IV and the linker are buried and indicate that the reorientation of helix III and opening of Structure 894

Figure 6. Formation of the Annexin I Binding Surface from Calcium Binding to Apo-S100A11 Accessible surface area representations of apo-S100A11 (A) and calcium-bound S100A11 (B) in complex with annexin I (Rety et al., 2000). Each protein is oriented approximately 90Њ with respect to that shown in Figure 4. In both molecules, residues in S100A11 that directly interact (Ͻ5A˚ ) with annexin I are indicated and colored dark blue and those whose side chain exposure increases by Ͼ20% upon calcium binding but have less interaction are colored cyan. For apo-S100A11 (A), many of these residues have less than 20% of their side chains accessible to the surface and are disjoint on the surface of the protein. In the calcium-bound structure (B), most residues that contact the annexin I peptide increase their side chain-accessible surface areas by Ͼ20% as shown in Figure 5C to form a contiguous site for the annexin interaction (dark blue). To facilitate comparison, all sequence numbers correspond to those of rabbit S100A11. the helix I-IV interaction must occur upon calcium bind- the apo-protein but is ␣ helical in the complex, along ing to position many of these residues for interaction with with F91, which also has contacts with annexin I. Muta- the annexin peptide. This results in a stunning transition genesis experiments with S100A11 have shown that the from a disjointed binding surface in apo-S100A11 to a region between residues E89 and Q97 is necessary for contiguous surface in the S100A11 complex (Figure 6). annexin I binding in the presence of calcium (Seemann The S100A11-annexin interaction (Rety et al., 2000) is et al., 1997). These observations indicate that the induc- characterized by the bridging of the annexin I peptide tion of helix in the A86–S90 region in apo-S100A11 upon with helix I of one monomer and helices III and IV and calcium binding is important for target protein interac- the linker of the second S100A11 monomer (Figures 4 tion. It is interesting that residues H85–F88 become heli- and 6). Interactions occur between P3 (S100A11; residue cal upon calcium binding in S100B also (Matsumura et numbering refers to the rabbit sequence) and A1 (annexin al., 1998; Smith and Shaw, 1998b). However, the similar I), and E11 (S100A11) and A1, M2, and V3 (annexin I), contacts found for S100B and S100A11 make it difficult including hydrogen bonds between the E11 side chain to determine how this region confers specificity at pres- carboxylate and M2 and V3 backbone amides. Similar ent. One difference may occur in the third of these posi- interactions are observed for S100A10-annexin II (Rety tions where S90 in S100A11 has close contacts (Ͻ4A˚ ) et al., 1999) but not for S100B, where interactions be- with F6 in annexin I (Figure 6) while F87 has more remote tween a p53 peptide or TRTK-12 are confined to individ- contacts (Ͼ5.0 A˚ ) with its bound targets in S100B. Fur- ual monomers (McClintock and Shaw, 2003; Rustandi ther, it is interesting that S100A11 is one of a handful et al., 2000). The positions of these residues in the se- of S100 proteins (S100A10, S100A3, and S100A4 are quences of S100 proteins are among the most variable others) that has an abnormally long C terminus, 10 and in S100 proteins. For example, P3 occupies a position 11 residues longer than S100B and S100A6, respec- that is absent in S100B and S100A12 due to the ex- tively. It is possible that induction of ␣ helix in this region tended N terminus in S100A11. The position occupied is especially important in these proteins to promote tar- by E11 in S100A11 is conserved in S100A10, S100A11, get interactions closer to the protein C terminus. In S100A12, and S100A9, but variable in S100B, S100A5, agreement with this, S100A10, which interacts with S100A6, and S100A3. Thus, it would appear that the annexin II, has been shown to contain numerous con- uncovering of residues P3 and E11 in apo-S100A11 by tacts from Y85, F86, and M90 near its C terminus (Rety opening up the helix I-IЈ interaction (i.e., the interhelical et al., 1999). It might be expected that S100A4 and angle) must play an important role for the bridging inter- S100A3 will have this similar feature once structures in actions of annexin I with this region. complex with calcium and a target protein are obtained. In helix IV, residues A86, C87, and S90 increase their In the linker of apo-S100A11, residues L43, A44, A45, exposure upon calcium binding and are required for F46, and D52 are between 50% and 100% buried (Figure target interactions (Figures 5 and 6). However, unlike 5). In the complex, these residues increase their expo- the differences noted in the C terminus of S100A11, sure and have important contacts with the annexin I these residues are less variable in the S100 sequences. peptide. The side chain of F46 in S100A11 has contacts Further, the identical positions in S100B (A83, C84, and with L7 and W11, and A44 is positioned nearby K8 in F87) possess interactions with the target peptides from annexin I. A45 increases its exposure by nearly 40% TRTK-12 and p53 (McClintock and Shaw, 2003; Rustandi from that in the apo-state and has contacts with S4, L7, et al., 2000). In S100A11, this region is unstructured in and K8 in annexin I (Figure 6). The sequence of S100A11 Structure of Apo-S100A11 895

in this region of the linker (LAAF) is significantly different by S100A11 would result in a conformational change from those of S100A10 (FPGF), S100B (LSHF), and S100A6 that promotes its rapid association with annexin I and (LTIϪ), especially for these central 2 residues. Further, stimulation of membrane association. all 4 residues in apo-S100A11 are more buried, become more exposed upon calcium binding and have important Experimental Procedures contacts with annexin I (Figures 5 and 6). In S100A10, it is the 2 “outside” residues (F38, F41) that interact with Sample Preparation Unlabeled and uniformly 15N- and 15N,13C-labeled rabbit S100A11 residues of annexin II (Rety et al., 1999). In S100B, there ˚ were expressed in Escherichia coli strain BL21(DE3) and purified as are no contacts within 5 A for residues L40–F43 with previously described (Rintala et al., 2002). Briefly, a 1 L culture was either p53 or TRTK-12 targets. In agreement with this, grown at 37ЊC with agitation to an OD600 of about 0.7–0.8 and then it has been shown that S100A11 does not interact with induced with 0.4 M IPTG for several hours. The cells were harvested the S100B target, TRTK-12 (Barber et al., 1999) or the by centrifugation, lysed using a French pressure cell, and then sub- Њ S100A10 target, annexin II (Rety et al., 1999). Further, jected to ultracentrifugation at 38,000 rpm and 4 C for 90 min. The supernatant was loaded onto a phenyl-Sepharose column with a sequence analysis reveals that residues in the linker calcium-containing buffer and then washed for several hours until region (including LAAF in S100A11) are among the most the OD280 returned to baseline. The protein was eluted from the column variable for all S100 sequences. These observations may using the same buffer containing EGTA instead of calcium. Fractions point to this sequence of residues being important for containing the protein were pooled, dialyzed, and lyophilized. Spe- the specificity of target proteins by individual S100 pro- cifically 15N-glycine/serine-labeled S100A11 was expressed in the teins, although further structural and biological experi- auxotrophic cell line, DL39GlyA, and grown on selective medium containing 15N-glycine as the sole glycine source. Fractionally mentation of other S100 proteins and their targets will 13C-labeled S100A11 was prepared by growing BL21(DE3) on M9 mini- be required to support this idea. mal medium containing a 1:10 mixture of fully 13C-labeled glucose and unlabeled glucose, respectively (Neri et al., 1989). Asymmetrically labeled S100A11 was prepared by incubating a 1:1 mixture of unla- Biological Implications beled S100A11 (36 ␮M) and 13C,15N-labeled S100A11 (36 ␮M) in 50 S100A11 is a member of the S100 protein family of EF- ml of water for approximately 220 hr in the presence of 0.2 mM Њ hand calcium binding proteins. The general mechanism dithiothreitol at 37 C with agitation. The progression of the reaction was monitored by electrospray mass spectrometry under nondena- for these proteins involves a calcium-induced conforma- turing conditions. The protein solution was lyophilized and then tional change allowing them to interact with other target dissolved in 5 ml of 100% D2O and incubated overnight and then proteins and elicit a biological response. To determine repeated. The solution was concentrated to 600 ␮l with a final mono- the details of this response, three-dimensional struc- mer concentration of 3 mM of each protein, deuterated dithiothreitol tures of the calcium-free, calcium-bound, and calcium- was added to a concentration of 10 mM, and the pH was adjusted bound with target protein are required. In this work, the to 7.25. All other NMR samples were prepared at a dimer concentration of 1 mM in 90% H O/10% D O (v/v), 50 mM KCl, 10 mM dithio- three-dimensional structure of apo-S100A11 has been 2 2 threitol (pH 7.25). determined using NMR spectroscopy. The structure reveals that the protein forms a symmetric dimer, with each NMR Spectroscopy monomer containing two EF hands. The conformation of NMR experiments were performed at 35ЊC on Varian 500, 600, and site I is similar to that recently determined in calcium- 800 MHz spectrometers with pulse field gradient triple-resonance bound S100A11 bound to an annexin I peptide, indicating probes. Sequential assignments of the polypeptide backbone re- that calcium binding to this site results in little reorganiza- sonances for apo-S100A11 were made from HNCA, HN(CO)CA, 1 15 tion. In contrast, helices III and IV, which surround cal- HNCACB, CBCA(CO)NH (Grzesiek and Bax, 1992), and H- N HSQC experiments (Kay et al., 1992). Side chain resonance assignments cium binding site II in apo-S100A11, are near antiparallel, were made from 15N-edited TOCSY, C(CO)NH, HCCH-TOCSY (Kay and helices I and IV are more closed than in other S100 et al., 1993), HC(CO)NH, HBCBCGCDHD, HBCBCGCDCEHE (Yama- proteins. Comparison to the calcium-bound S100A11 zaki et al., 1993), and 1H-13C HSQC experiments. The 1H, 13C, and structure shows that helix I-IV and I-IЈ interactions open 15N resonances of apo-S100A11 have been assigned and deposited slightly, while the helix III-IV arrangement is dramatically in the BioMagResBank under accession number BMRB-5189. Frac- 13 13 altered by nearly 40Њ. Together, these structural changes tionally C-labeled S100A11 protein was used to collect C-coupled and decoupled 1H-13C HSQC spectra in order to stereo-specifically unmask key residues and provide a contiguous binding assign prochiral methyl groups for valine and leucine residues. The surface for annexin I target binding. 13C-coupled 1H-13C HSQC was collected using 512 increments in One of the proposed biological functions for S100A11 order to be able to observe the splitting in the 13C dimension. Amide is to aid in the aggregation of membrane surfaces impor- exchange was determined by dissolving S100A11 protein in 100% 1 15 tant for cell vesiculation. Annexin I has been shown to D2O and collecting a H- N HSQC spectrum after a 10 hr incubation. 15 13 interact with phospholipids in a calcium-sensitive man- Interproton distances were obtained from N-NOESY and C, 15N-NOESY spectra. The 15N-NOESY spectrum was collected on ner. Further, interaction between S100A11 and annexin a Varian 600 MHz spectrometer with a 150 ms mixing time. The I is dependent on calcium binding to S100A11. Such a 13C,15N-NOESY spectrum (Pascal et al., 1994) was collected on an 13 calcium-stimulated mechanism would require tight reg- INOVA 800 MHz spectrometer with a 100 ms mixing time. A CF1- ulation by the calcium binding protein S100A11 in order filtered, F3-edited NOESY experiment (Zwahlen et al., 1997) was to time and control membrane aggregation events. The collected on an INOVA 800 MHz spectrometer with a 150 ms mixing structure of apo-S100A11 reveals that residues impor- time. tant for the interaction with annexin I are hidden from its surface, therefore minimizing the contact between Structure Determination Structures of apo-S100A11 were calculated using the simulated the two proteins. This would allow S100A11 and annexin annealing protocol in the CNS program (Brunger et al., 1998). Di- I to be closely located together near the membrane with meric S100A11 structures were generated using a total of 2964 little effect on membrane association. Calcium binding nOes, including 1336 intraresidue, 658 sequential, 594 short-range, Structure 896

264 long-range, and 112 intermonomer distances as well as 64 (1992). S-100 protein binds to annexin II and p11, the heavy and hydrogen bond distance restraints and 192 dihedral restraints. In- light chains of calpactin I. Biochim. Biophys. Acta 1160, 67–75. terproton distances for all proton pairs, except intermonomer nOes, Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., were calibrated using maximum and minimum nOe intensities for Grosse-Kunstleve, R.W., Jiang, J.-S., Kunzewski, J., Nilges, M., known possible dNH␣ distances. Dihedral angle restraints were deter- Pannu, N.S., et al. (1998). Crystallography & NMR system: a new mined by the TALOS (torsion angle likelihood obtained from shifts software suite for macromolecular structure determination. Acta and sequence similarity) program (Cornilescu et al., 1999) based on Crystallogr. D54, 905–921. chemical shift data from the H␣,C␣,C␤, and CO resonances of Cornilescu, G., Delaglio, F., and Bax, A. (1999). Protein backbone apo-S100A11 and sequence similarity. Angles were selected when angle restraints from searching a database for chemical shift and nine of ten dihedral matches fell within an allowed region of the sequence homology. J. Biomol. NMR 13, 289–302. Ramachandran plot. These angles were restricted to ␺Ϯerror Њ and φ Ϯ error Њ for structure calculations where the error was the Delphin, C., Ronjat, M., Deloulme, J.C., Garin, G., Debussche, L., deviation from the average angle selected by TALOS. Hydrogen Higashimoto, Y., Sakaguchi, K., and Baudier, J. (1999). Calcium- bonds were identified from slowly exchanging NH resonances after dependent interaction of S100B with the C-terminal domain of the tumor suppressor p53. J. Biol. Chem. 274, 10539–10544. a 10 hr incubation of apo-S100A11 in D2O. For each hydrogen bond, two distance restraints were used, NH-O (1.8–2.3 A˚ ) and N-O Donato, R. (2001). S100: a multigenic family of calcium-modulated (2.3–3.3 A˚ ). Structure determination of the apo-S100A11 dimer in- proteins of the EF-hand type with intracellular and extracellular func- volved the use of noncrystallographic symmetry. Intermonomer dis- tional roles. Int. J. Biochem. Cell Biol. 33, 637–668. tance restraints were not calibrated and were set to maximum and Drohat, A.C., Baldisseri, D.M., Rustandi, R.R., and Weber, D.J. ˚ minimum distances of 6.00 and 1.70 A, respectively, for all calculations. (1998). Solution structure of calcium-bound rat S100B (␤␤) as determined by nuclear magnetic resonance spectroscopy. Biochemistry Acknowledgments 37, 2729–2740. Drohat, A.C., Tjandra, N., Baldisseri, D.M., and Weber, D.J. (1999). The authors would like to thank Kathryn Barber (UWO) for her excel- The use of dipolar couplings for determining the solution structure lent technical support, Frank Delaglio and Dan Garrett (NIH) for the of rat apo-S100B(␤␤). Protein Sci. 8, 800–809. programs NMRPipe and Pipp, Lewis Kay (University of Toronto) for all pulse sequences, and Brett Scho¨ nekess (University of Calgary) Fritz, G., Mittl, P.R.E., Vasak, M., Grutters, M.G., and Heizmann, for the S100A11 cDNA used in this work. We would like to thank C.W. (2002). The crystal structure of metal-free human EF-hand the Canadian National High Field NMR Centre (NANUC, Edmonton, protein S100A3 at 1.7A˚ resolution. J. Biol. Chem. 277, 33092–33098. 13 Alberta) for acquisition of the CF1-edited, F3-filtered NOESY-HSQC Fujii, T., Machino, K., Andoh, H., Satoh, T., and Kondo, Y. (1990). experiment and Dr. L. Spyracopoulos (University of Alberta) for help- Calcium-dependent control of caldesmon-actin interaction by S100 ful discussions. This research was supported from operating and protein. Biochemistry 107, 133–137. maintenance grants from the Canadian Institutes of Health Research Garbuglia, M., Verzini, M., and Donato, R. (1998). Annexin VI binds (GSS) and graduate studentships from the Natural Sciences and S100A1 and S100B and blocks the ability of S100A1 and S100B to Engineering Research Council and Canadian Institutes of Health inhibit desmin and GFAP assemblies into intermediate filaments. Research (A.C.D.). M.P.W. is an Alberta Heritage Foundation for Cell Calcium 24, 177–191. Medical Research Scientist and a recipient of a Canada Research Chair (Tier I) in Biochemistry. The 600 MHz NMR spectrometer at Garbuglia, M., Verzini, M., Rustandi, R.R., Osterloh, D., Weber, D.J., UWO was funded through grants from the Canada Foundation for Gerke, V., and Donato, R. (1999). Role of the C-terminal extension Innovation, Ontario Innovation Trust, and the Academic Develop- in the interaction of S100A1 with GFAP, tubulin, the S100A1- and ment Fund of The University of Western Ontario. Operation of NA- S100B-inhibitory peptide, TRTK-12, and a peptide derived from p53, NUC is funded by the Canadian Institutes of Health Research, the and the S100A1 inhibitory effect on GFAP polymerization. Biochem. Natural Sciences and Engineering Research Council of Canada, and Biophys. Res. Commun. 254, 36–41. the University of Alberta. Gerke, V. (1990). Tyrosine kinase substrate annexin II (p36)— biochemical characterization and conservation among species. Bio- Received: January 31, 2003 chem. Soc. Trans. 18, 1106–1108. Revised: April 4, 2003 Gerke, V., and Moss, S.E. (2002). Annexins: from structure to func- Accepted: April 18, 2003 tion. Physiol. Rev. 82, 331–371. Published: July 1, 2003 Grzesiek, S., and Bax, A. (1992). An efficient experiment for sequential backbone assignment of medium-sized isotopically enriched References proteins. J. Magn. Reson. 99, 201–207. Heizmann, C.W. (1999). Ca2ϩ-binding S100 proteins in the central Akke, M., Drakenberg, T., and Chazin, W. (1992). Three-dimensional 2ϩ nervous system. Neurochem. Res. 24, 1097–1100. solution structure of Ca loaded porcine calbindin D9k determined by nuclear magnetic resonance spectroscopy. Biochemistry 31, Heizmann, C.W., Fritz, G., and Schafer, B.W. (2002). S100 proteins: 1011–1020. structure, functions and pathology. Front. Biosci. 7, d1356–d1368. Allen, B.G., Durussel, I., Walsh, M.P., and Cox, J.A. (1996). Charac- Ikura, M. (1996). Calcium binding and conformational response in terization of the Ca2ϩ-binding properties of calgizzarin (S100C) iso- EF-hand proteins. Trends Biochem. Sci. 21, 14–17. lated from chicken gizzard smooth muscle. Biochem. Cell Biol. 74, Ikura, M., Clore, G.M., Gronenborn, A.M., Zhu, G., Klee, C.B., and 687–694. Bax, A. (1992). Solution structure of a calmodulin-target peptide Barber, K.R., McClintock, K.A., Jamieson, G.A., Jr., Dimlich, R.V., complex by multidimensional NMR. Science 256, 632–638. ϩ and Shaw, G.S. (1999). Specificity and Zn2 enhancement of the Kay, L.E., Keifer, P., and Saarinen, T. (1992). Pure absorption gradi- S100B binding epitope TRTK-12. J. Biol. Chem. 274, 1502–1508. ent enhanced heteronuclear single quantum correlation spectros- Baudier, J., Delphin, C., Grunwald, D., Khochbin, S., and Lawrence, copy with improved sensitivity. J. Am. Chem. Soc. 114, 10663– J.J. (1992). Characterization of the tumor suppressor protein p53 10665. as a protein kinase C substrate and a S100b-binding protein. Proc. Kay, L.E., Xu, G., Singer, A.U., Muhandiram, D.R., and Forman-Kay, Natl. Acad. Sci. USA 89, 11627–11631. J.D. (1993). A gradient-enhanced HCCH-TOCSY experiment for re- 1 13 Becker, T., Gerke, V., Kube, E., and Weber, K. (1992). S100P, a cording side-chain H and C correlations in H2O samples of pro- novel Ca2ϩ-binding protein from human placenta: cDNA cloning, teins. J. Magn. Reson. 101, 333–337. 2ϩ recombinant protein expression and Ca binding properties. Eur. Kilby, P.M., Van Eldik, L.J., and Roberts, G.C.K. (1996). The solution J. Biochem. 207, 541–547. structure of the bovine S100b protein dimer in the calcium-free Bianchi, R., Pula, G., Ceccarelli, P., Giambanco, I., and Donato, R. state. Structure 4, 1041–1052. Structure of Apo-S100A11 897

Kordel, J., Skelton, N.J., Akke, M., and Chazin, W.J. (1993). High Hidaka, H., and Chazin, W.J. (1998). The three-dimensional structure 2ϩ 2ϩ resolution solution structure of calcium-loaded calbindin D9k. J. Mol. of Ca -bound calcyclin: implications for Ca -signal transduction Biol. 231, 711–734. by S100 proteins. Structure 6, 223–231. Maler, L., Potts, B.C., and Chazin, W.J. (1999). High resolution solu- Scho¨ nekess, B.O., and Walsh, M.P. (1997). Molecular cloning and tion structure of apo calcyclin and structural variations in the S100 expression of avian smooth muscle S100A11 (calgizzarin, S100C). family of calcium-binding proteins. J. Biomol. NMR 13, 233–247. Biochem. Cell Biol. 75, 771–775. Maler, L., Sastry, M., and Chazin, W.J. (2002). A structural basis for Seemann, J., Weber, K., and Gerke, V. (1996). Structural require- S100 protein specificity derived from comparative analysis of apo ments for annexin I-S100C complex formation. Biochem. J. 319, and Ca2ϩ-calcyclin. J. Mol. Biol. 317, 279–290. 123–129. Martin, S.R., Bayley, P.M., Brown, S.E., Porumb, S.E., Zhang, M., and Seemann, J., Weber, K., and Gerke, V. (1997). Annexin I targets Ikura, M. (1996). Spectroscopic characterization of a high-affinity S100C to early endosomes. FEBS Lett. 413, 185–190. calmodulin-target peptide hybrid molecule. Biochemistry 35, 3508– Skelton, N.J., Kordel, J., and Chazin, W.J. (1995). Determination of 3517. the solution structure of apo calbindin D9k by NMR spectroscopy. Matsumura, H., Shiba, T., Inoue, T., Harada, S., and Kai, Y. (1998). J. Mol. Biol. 249, 441–462. ˚ A novel mode of target recognition suggested by the 2.0A structure Smith, S.P., and Shaw, G.S. (1998a). A change-in-hand mechanism of holo S100B from bovine brain. Structure 6, 233–241. for S100 signalling. Biochem. Cell Biol. 76, 324–333. McClintock, K.A., and Shaw, G.S. (2003). A novel S100 target confor- Smith, S.P., and Shaw, G.S. (1998b). A novel calcium-sensitive 2ϩ mation is revealed by the solution structure of the Ca -S100B- switch revealed by the structure of human S100B in the calcium- TRTK-12 complex. J. Biol. Chem. 278, 6251–6257. bound form. Structure 6, 211–222. McKay, R.T., Tripet, B.P., Hodges, R.S., and Sykes, B.D. (1997). Vallely, K.M., Rustandi, R.R., Ellis, K.C., Varlamova, O., Bresnick, Interaction of the second binding region of troponin I with the regula- A.R., and Weber, D.J. (2002). Solution structure of human Mts1 tory domain of skeletal muscle troponin C as determined by NMR (S100A4) as determined by NMR spectroscopy. Biochemistry 41, spectroscopy. J. Biol. Chem. 272, 28494–28500. 12670–12680. Neri, D., Szyperski, T., Otting, G., Senn, H., and Wuthrich, K. (1989). Van Eldik, L.J., and Griffin, W.S.T. (1994). S100b expression in Alzhei- Stereospecific nuclear magnetic resonance assignments of the mer’s disease: relation to neuropathology in brain regions. Biochim. methyl groups of valine and leucine in the DNA-binding domain of Biophys. Acta 1223, 398–403. the 434 repressor by biosynthetically directed fractional 13C labeling. Wimberly, B., Thulin, E., and Chazin, W.J. (1995). Characterization Biochemistry 28, 7510–7516. of the N-terminal half-saturated state of calbindin D9k: NMR studies Odink, K., Cerletti, N., Bruggen, J., Clerc, R.G., Tarcsay, L., Zwadlo, of the N56A mutant. Protein Sci. 4, 1045–1055. G., Gerhards, G., Schlegel, R., and Sorg, C. (1987). Two calcium- Yamazaki, T., Forman-Kay, J.D., and Kay, L.E. (1993). Two-dimen- binding proteins in infiltrate macrophages of rheumatoid arthritis. sional NMR experiments for correlating 13C␤ and 1H␦/⑀ chemical Nature 330, 80–82. shifts of aromatic residues in 13C-labeled proteins via scalar cou- Olwin, B.B., and Storm, D.R. (1985). Calcium binding to complexes plings. J. Am. Chem. Soc. 115, 11054–11055. of calmodulin and calmodulin binding proteins. Biochemistry 24, Yazawa, M., Ikura, M., Hikichi, K., Ying, L., and Yagi, K. (1987). 8081–8086. Communication between two globular domains of calmodulin in the Otterbein, L.R., Kordowska, J., Witte-Hoffmann, C., Wang, C.L.A., presence of mastoparan or caldesmon fragment. Ca2ϩ and 1H NMR. 2ϩ and Dominguez, R. (2002). Crystal structures of S100A6 in the Ca - J. Biol. Chem. 262, 10951–10954. free and Ca2ϩ-bound states: the calcium sensor mechanism of S100 Zhao, X.Q., Naka, M., Muneyuki, M., and Tanaka, T. (2000). Ca2ϩ- proteins revealed at atomic resolution. Structure 10, 557–567. dependent inhibition of actin-activated myosin ATPase activity by Pascal, S.M., Muhandiram, D.R., Yamazaki, T., Forman-Kay, J.D., S100C (S100A11), a novel member of the S100 protein family. Bio- and Kay, L.E. (1994). Simultaneous acquisition of 15N- and 13C-edited chem. Biophys. Res. Commun. 267, 77–79. NOE spectra of protein dissolved in H O. J. Magn. Reson. B 103, 2 Zwahlen, C., Legault, P., Vincent, S.J.F., Greenblatt, J., Konrat, R., 197–201. and Kay, L.E. (1997). Methods for measurement of intermolecular Petrova, T.V., Hu, J., and Van Eldik, L.J. (2000). Modulation of glial NOEs by multinuclear NMR spectroscopy: application to a bacterio- activation by astrocyte-derived protein S100B: differential re- phage ␭ N-peptide/boxB RNA complex. J. Am. Chem. Soc. 119, sponses of astrocyte and microglial cultures. Brain Res. 853, 74–80. 6711–6721. Potts, B.C.M., Smith, J., Akke, M., Macke, T.J., Okazaki, K., Hidaka, H., Case, D.A., and Chazin, W.J. (1995). The structure of calcyclin Accession Numbers reveals a novel homodimeric fold for S100 Ca2ϩ-binding proteins. Nat. Struct. Biol. 2, 790–796. The coordinates for apo-S100A11 have been deposited in the Pro- Rety, S., Sopkova, J., Renouard, M., Osterloh, D., Gerke, V., Tabar- tein Data Bank under ID code 1NSH. ies, S., Russo-Marie, R., and Lewit-Bentley, A. (1999). The crystal structure of a complex of p11 with the annexin II N-terminal peptide. Nat. Struct. Biol. 6, 89–95. Rety, S., Osterloh, D., Arie, J.-P., Tabaries, S., Seemann, J., Russo- Marie, F., Gerke, V., and Lewit-Bentley, A. (2000). Structural basis of the Ca2ϩ-dependent association between S100C (S100A11) and its target, the N-terminal part of annexin I. Structure 8, 175–184. Rintala, A.C., Scho¨ nekess, B.O., Walsh, M.P., and Shaw, G.S. (2002). 1H, 15N and 13C resonance assignments of rabbit apo-S100A11. J. Biomol. NMR 22, 191–192. Rustandi, R.R., Baldisseri, D.M., and Weber, D.J. (2000). Structure of the negative regulatory domain of p53 bound to S100B. Nat. Struct. Biol. 7, 570–574. Rustandi, R.R., Baldisseri, D.M., Inman, K.G., Nizner, P., Hamilton, S.M., Landar, A., Zimmer, D.B., and Weber, D.J. (2002). Three-dimensional solution structure of the calcium-signaling protein apo- S100A1 as determined by NMR. Biochemistry 41, 788–796. Sastry, M., Ketchem, R.R., Crescenzi, O., Weber, C., Lubienski, M.J.,